Keynote: Distributed Computing 4 Kids -- with Spark
summary
A concrete claim: kids can learn distributed systems concepts first and syntax later by mapping work to garden gnomes and using Spark as a teaching vehicle. The materials include a PDF book with gnome metaphors, a Python follow‑up with pip install examples, and notes showing why the shuffle is painfully slow and leader coordination avoids doing worker work. Alternatives discussed: Dask and Ray for task/actor models. Illustrations are unfinished but the Python demos exist.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:07Meet Holden Karau
- 3:52Teaching distributed computing
- 5:36Introducing Spark and FP
- 7:01Concepts before syntax
- 8:27Why it matters for kids
- 10:36Garden gnomes and Instagram
- 14:50Leader election explained
- 17:00Recording and combining results
- 19:58Handling failures and shuffle
- 22:05From gnomes to Python
- 24:54Future plans for parallelism
- 27:44Employer message
- 28:49Inviting collaboration
- 29:32Q&A: Lesson plans
- 32:03Q&A: Spark vs Dask vs Ray
- 37:06Q&A: Global gnome scaling
- 39:53Q&A: Spark and Santa
- 40:36Q&A: Explaining to stakeholders
- 43:26Closing
related talks
42mScala: The Unpredicted Lingua Franca for Data Science
26mOptimal Approaches for Real-Time Machine Learning with Apache Spark on Kubernetes.
01h 00mGetting Started with Spark
39mScaling Python for Machine Learning: Beyond Data Parallelism
44mData Science at Scale with Apache Spark
29mTed Willke, Senior Principal Engineer & GM, Datacenter Group, Intel
01h 07mKeynote + Conversation
02h 18mHacking Dask: Diving Into Dask;s Internals